虚拟人火了，你准备好了吗？
AI虚拟人算法整理汇总
- 一、虚拟人脸
-
- 1. StyleGan
- 2. StyleGan2
- 3. StyleGan2-ada
- 4. StyleGan3
- 二、人脸同步
-
- 1. First Order Model
- 2. Avatarify
- 3. PC-AVS: Talking Face Generation
- 三、AI换脸
-
- 1. FaceSwap
- 2. DeepFaceLab
- 3. DeepFaceLive
- 四、嘴型生成
-
- 1. LipGan
- 2. Wav2Lip
- 3. MakeItTalk：
从2019年开始接触虚拟主播到现在，各个方向的项目多多少少都做过一遍。近期打算整理虚拟人相关算法资料，会不定时更新，敬请关注。虚拟人算法相关分类与应用主要有这几种： 虚拟人脸（人脸生成）、 人脸同步以及 AI换脸。
Ayayi、阿喜、柳夜熙这种偏向建模类型的虚拟人会另外开贴介绍，敬请期待。
一、虚拟人脸
代表算法有 Nvidia Lab 研究发布的 StyleGan 系列开源代码。
StyleGan 是由英伟达实验室发布的虚拟人脸生成算法，可以无中生有，生成非常逼真的虚拟人(数字人)。AI通过学习大量的高清人脸图片之后，就能生成几十亿现实世界中不存在的人脸。
1. StyleGan
A Style-Based Generator Architecture for Generative Adversarial Networks (CVPR 2019)
Github: https://github.com/NVlabs/stylegan
These people are not real – they were produced by our generator that allows control over different aspects of the image.
2. StyleGan2
Analyzing and Improving the Image Quality of StyleGAN (CVPR 2020)
Github: https://github.com/NVlabs/stylegan2
3. StyleGan2-ada
Training Generative Adversarial Networks with Limited Data (CVPR 2020)
Github: https://github.com/NVlabs/stylegan2-ada-pytorch
4. StyleGan3
Alias-Free Generative Adversarial Networks (CVPR 2021)
Github: https://github.com/NVlabs/stylegan3
stylegan 出现的时候已经让人惊掉下巴了，第三代更加细腻逼真。具体改进可以看github主页。
二、人脸同步
1. First Order Model
First Order Motion Model for Image Animation (NeurIPS 2019)
Github: https://github.com/AliaksandrSiarohin/first-order-model
2. Avatarify
Photorealistic avatars for video-conferencing.
Github: https://github.com/alievk/avatarify
3. PC-AVS: Talking Face Generation
Pose-Controllable Talking Face Generation by Implicitly Modularized Audio-Visual Representation (CVPR 2021)
Github: https://github.com/Hangz-nju-cuhk/Talking-Face_PC-AVS
三、AI换脸
1. FaceSwap
FaceSwap is a tool that utilizes deep learning to recognize and swap faces in pictures and videos.
Github: https://github.com/deepfakes/faceswap
Jennifer Lawrence/Steve Buscemi FaceSwap using the Villain model
2. DeepFaceLab
DeepFaceLab: Integrated, flexible and extensible face-swapping framework (CVPR 2021)
Github: https://github.com/iperov/DeepFaceLab
3. DeepFaceLive
Github: https://github.com/iperov/DeepFaceLive
四、嘴型生成
1. LipGan
Towards Automatic Face-to-Face Translation (CVPR 2020)
Github: https://github.com/Rudrabha/LipGAN
2. Wav2Lip
A Lip Sync Expert Is All You Need for Speech to Lip Generation In The Wild (CVPR 2020)
Github: https://github.com/Rudrabha/Wav2Lip
3. MakeItTalk：
MakeItTalk: Speaker-Aware Talking-Head Animation
Github: https://github.com/yzhou359/MakeItTalk
